Files in this item



application/pdf9124377.pdf (4MB)Restricted to U of Illinois
(no description provided)PDF


Title:Kernel method in Monte Carlo importance sampling
Author(s):Ang, George Lee
Doctoral Committee Chair(s):Tang, Wilson H.
Department / Program:Civil and Environmental Engineering
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Engineering, Civil
Abstract:A new approach to evaluate the reliability of structural systems using a Monte Carlo variance reduction technique called the Importance Sampling is presented. Since the efficiency of the importance sampling method depends primarily on the choice of the importance sampling density, the use of the kernel method to estimate the optimal importance sampling density is proposed.
The first step in implementing the proposed method involves generating samples from the original distribution. The kernel sampling density is then constructed using these samples. A second set of samples is then generated from the kernel sampling density, and the failure probability is estimated by taking the average of the two sets of samples.
A number of example problems were examined to illustrate the application of the proposed kernel method. The method was shown to be more efficient than the basic Monte Carlo method and yielded unbiased probability of failure estimates. It was demonstrated to perform better than the adaptive sampling method. The method was also shown to be versatile because it can be applied to problems with very complex performance functions that can not be expressible in explicit form, and to produce unbiased estimate of the failure probability even in problems with multiple failure modes. In problems with large number of random variables, the efficiency of the kernel method increased after treating the unimportant random variables as constants.
Issue Date:1991
Rights Information:Copyright 1991 Ang, George Lee
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9124377
OCLC Identifier:(UMI)AAI9124377

This item appears in the following Collection(s)

Item Statistics